Sensitive Determination of PAH-DNA Adducts

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Matrix Design for Matrix-Assisted Laser Desorption Ionization: Sensitive Determination of PAH-DNA Adducts

M. George, J. M. Y. Wellemans, R. L. Cerny, and M. L. Gross* Midwest Center for Mass Spectrometry, Department of Chemistry, University of Nebraska-Lincoln, Lincoln, Nebraska USA

K. Li and E. L. Cavalieri Eppley institute for Research in Cancer and Allied Diseases, University of Nebraska Medical Center, 600 South 42nd Street, Omaha, Nebraska USA

Two matrices, 4-phenyl-cw-cyanocinnamic acid (ICC) and 4-benzyloxy-cu-cyanocinnamic acid (BCC), were identified for the determination of polycyclic aromatic hydrocarbon @‘AH) adducts of DNA bases by matrix-assisted laser desorption ionization. These matrices were designed based on the concept that the matrix and the analyte should have structural similarities. PCC is a good matrix for the desorption of not only PA&modified DNA bases, but also PAHs themselves and their rnetabolites. Detections can be made at the femtomolar level. (1 Am Sot Mass Spectrum 1994, 5, 1021-1025)

he success of matrix-assisted laser desorption quire sensitive detection methods at the low picomole ionization (MALDI) for the sensitive detection of and mid-femtomole level, respectively [6]. T an analyte depends on the nature of the matrix. One method for detection that also gives structural Rational design of a matrix requires understanding of information is fast-atom bombardment (FAB) coupled the mechanism of MALDI. The mechanism is not fully with tandem mass spectrometry [6-81. Unfortunately, understood. Nevertheless, empirical criteria for matrix the combination lacks sensitivity because there is in selection have been suggested 111. Two important ones, efficiency of FAB desorption and interference from in our view, are (1) high molar absorptivity and (21 matrix-derived ions. Improvements can be made by miscibility with the analyte in the solid phase. derivatizing the modified base 191 and by using coaxial The question to be answered is how does one assure continuous-flow FAB [lo]. miscibility in the solid state? We propose that miscibil- MALDI ionization may offer high sensitivity for ity is best achieved when the analyte and matrix modified DNA bases as it has for peptides Ill] and molecules have similar structures. In this paper, we proteins [12J. Encouragement comes from the success- expand upon the guidelines proposed earlier [l] and ful application of MALDI coupled with time-of-flight describe the design and evaluation of a matrix molecule mass spectrometry to other molecules such as poly- for the specific problem of desorbing polycyclic aro- mers [13-151 and oligosaccharides [16]. For DNA and matic hydrocarbon (PAHI adducts of DNA bases. Our RNA, success has been realized with both large [17] argument at this stage is qualitative, and we do not and small [I81 oligonucleotides, nucleotides [19], and address quantitatively the solubility of an analyte in one modified nucleoside, which was detected at the the matrix. femtomolar level [20]. MALDI coupled with Fourier Some PAH materials are carcinogenic [Zl because transform mass spectrometry also has been demon- they are metabolically activated to form electrophiles strated for the determination of modified nucleosides that attack nucleophilic sites on DNA bases. Three produced by attack of PAH diol epoxides on DNA, but mechanisms of activation of PAHs have been pro- the required quantities of material thus far are in the posed: (1) formation of dioI epoxide [3], (2) formation 20-pm01 range [Zl, 221. of a radical cation [4], and (3) formation of benzylic esters [5]. Investigations of the relative importance of these mechanisms in in vitro and in vivo studies re- Experimental

The synthesis of the PAH-modified DNA bases was Address reprint requests to Professor Michael L. Gross, Editor-in- published elsewhere [23]. Siphenyl carboxylic acid, Chief, Journal of Mass Spectromctry, Department of Chemishy, Washington University. 1 Brooking Drive, St. Louis, MO 63130. anthracene-9-carboxylic acid, and 4-hydroxy-o-cyan*

8 1994 American Society for Mass Spectrometry Received June 9,1994 1044~0305/94/$7.00 Revised July 13,1994 Accepted August 8,1994 1022 GEORGE ET AL. J Am Sac Mass Spectram 1994,5,1021-1025

cinnamic acid (4HCC) were purchased from Aldrich 5.26 (s, lH), 6 7.27 (d, 2H), S 7.38-7.43 (m, 3H), 6 Chemical Co. (St. Louis, MO). 4-Benzyloxy-o-cyanocin- 7.50-7.52 Cd, 2H), S 8.09-8.12 Cd, 2H), 6 8.24 ts, lH), namic acid (BCC) (mp 204205 ‘C) was synthesized by for BCC] with a GE Omega spectrometer (General reacting benzyl bromide (1.21 g) with 4HCC (0.9 g) in Electric, Fremont, CA) operating at 300 MHz. Nega- methanol. Potassium hydroxide (0.8 g) was added to tive-ion FAB that used 3nitrobenzyl alcohol as matrix deprotonate the 4HCC. 4-Phenyl-cu-cyanocinnamic acid produced abundant [M - HI- ions of m / z 278.0821 (KC) was synthesized following the literature proce (calculated for C,,H,,NO,, 278.0817) and 248.0710 dure [24], mp 243-244 “C (lit. [24] mp 243-245 “C; see (calculated for C,,H,,NO,, 248.0712) for BCC and structures). KC, respectively. The compounds were pure at least as could be determined by thin-layer chromatography and melting point. The spectra were obtained with a

COOH Kratos MS-50 three-sector mass spectrometer (Kratos Analytical, Ramsey, NJ) equipped with an argon FAB

@$Q o-o-H gun. The MALDI-TOF experiments were carried out on a

Anthracene-9.carboxylic acid ~ipheny~carboxyli~ acid BenchTOF Bruker instrument (Bruker-Franzen Ana- lytik GMBH, Bremen, Germany). A nitrogen laser beam (337 nm, 20-kW peak laser power, spot size of 1 mm*, and 1.25-ns pulse width) was used to desorb the sam- HO CH=C c @@H-Cc -o- ples. The ions were accelerated with a potential of 25 kV.

4HCC PCC Samples of benzo[ a]pyrene-6-N7-guanine (BP&N7- Gua), 1 [25], dibenzd a,f]pyrene-lo-N7-guanine (DBP- N7-Gua), 2 1261, dibenzo[ a,l]pyrene-IO-N7-adenine [26],and dibenzo[ a,l]pyrene-8,9-diol (DBP-8,9-diol), 3 [27] were obtained from synthesis as previously de- scribed. Dibenzo[a,I]pyrene-11,12-diol (DBP-11,lZ ECC diol), 4, was purchased from CHBMSYN Science Labo- ratories (Lenexa, KS), benzotelpyrene (B[e]P), 5, was purchased from Sigma Chemical Co. (St. Louis, MO), and dibenzoJa,l]pyrene (DBP), 6, was obtained from the National Cancer Institute Chemical Carcinogen Repository (Bethesda, MD) (see structures). Stock solutions (11 pmol/pL for BP-6-N7Gua, 2.5 pmol/FL for DBP-N’-Gua, 100 nmol/pL for dibenzo[ a,Z]pyrene-lO- N7-adenine, 10 pmol/pL for DBP-8,9-dial, 10 pmol/pL for DBP-ll,IZ-diol, I nmol/l_~L for B[e]P, and 10 pmol/gL for DBP) were made by dissolving the analyte in methanol. Saturated solutions of the matrices were made in a mixture (40:60) of acetonitrile and water (1% trifluoroacetic acid). Samples were prepared by mixing 1 PL of matrix and variable volumes (up to 1 FL) of analyte on the probe tip and allowing them to crystallize prior to irradiation with the laser. It was not possible to use calibration plots to deter- mine the limits of detection because the signals produced by MALDI were highly variable. As an expedi- ent, the lowest amounts of analytes required for repro- ducible observation of their molecular ions (signal-to- noise ratio of at least 3:l) for three sample loadings were taken as detection limits. These limits were estab- lished by loading progressively lower and lower amounts of analyte. Fifty to sixty laser shots were signal-averaged for each spectrum. A blank of each a 4 matrix was run to identify the matrix peaks that might PCC and BCC were characterized by ‘H NMR JO confound an analyte peak. Data acquisition was con- 7.40-7.55 (m, 3H), S 7.75-7.80 (d, 2H), 6 7.87-7.95 (d, trolled by using the Bruker SUN data system and 2H), S 8.17-8.22 (d, 2H), S 8.37 (s, 1H) for KC and S XTOF software. J Am Sot Mass Spectrom 1994, 5,1021-1025 MATRIX DESIGN FOR MALDI 1023

Results and Discussion (Met-i)’ The structures of the PAH-modified DNA bases (for I 402 example benzo[ alpyrene-6-N7-guanine (BP&N7-Gua), 1, and dibenzo[ a,I]pyrene-lo-N’-guanine (DBP-N7- Gus), 2 have nonpolar aromatic ring systems on one end and a polar nitrogen heterocycle on the other. In an attempt to match the nature of the matrix with that of the analyte, we selected a set of aromatic compounds-biphenyl carboxylic acid, anthracene-9- carboxylic acid, 4-benzyloxy-ct-cyanocinnamic acid (BCC), and 4-phenyl-a-cyanocinnamic acid (PCC)--as Figure 1. MALDI-TOF mass spectrum of 11 pmol of BP-6-N7- candidate matrices for this class of PAH-modified Gus obtained by using the 4HCC matrix. The [M + HI+ ion is of m/z 402. The asterisks (‘1 indicate matrix ions. bases. These matrix candidates are similar in structure to the adducts because they have a polar group on one end of the molecule and aromatic rings on the other. The matrix, 4-hydroxy-a-cyanocinnamic acid (4HCC1, to give a high abundance of radical cations when one of the common MALDI matrices used for desorb- biphenyl carboxylic acid was the matrix. A modified ing peptides, is included in this study for comparison. base, dibenzot a,l]pyrene-lo-N7-adenine, however, Note that 4HCC is a molecule that has polar groups on could not be detected for a lo-pmol loading. Lack of both ends, and it has been used with good success to detectability of the modified base may be due to the detect peptides, which possess a number of polar low molar absorptivity for the biphenyl carboxylic acid groups, at the femtomolar level. Metabolites such as (~IIWX= 265 nm) at the wavelength of the incident dibenzo[ a,l]pyrene-8,9-diol (DBP-8,9-dial), 3, and laser beam (337 nm), whereas the radical cation is dibenzc$a,l]pyrenc-11,lZdiol (DBP-11,12-dial), 4, and detected readily because it desorbs directly (i.e., with- the PAHs, benzo[ elpyrene (B[e]P>, 5, and dibenzo out matrix assistance). [a,l]pyrene (DBP), 6, do not contain the polar DNA Anthracene-9-carboxylic acid, another molecule that base, and an electrophilic matrix, which may promote contains aromatic rings and has two strong UV absorp- radical cation formation, may be indicated for these tion bands near 337 nm (327 and 347 run) gave effec- substances. tive desorption of both adducts and metabolites. The We have observed that PAH-modified bases do limits of detection are, at best, 10 pmol. The rigid ring desorb from 4-hydroxy-a-cyanocinnamic acid (4HCC) system of anthracene-9-carboxylic acid may prevent matrix. This matrix, however, does not provide detec- optimal crystallization. tion limits low enough for analyzing samples for in On the basis of the foregoing observations, we vitro studies. The MALDJ-TOF mass spectrum of sought a better MALDI matrix for PAH metabolites benzo[a]pyrene-6-N7-guani (BP-6-N7-Gus) adduct and PAH-DNA adducts. Two candidates were C-ben- (see Figure 1) obtained by using 11 pmol of the ana- zyloxy-a-cyanocinnamic acid (BCC) and 4-phenyl-a- lyte, shows an abundant [M + HI+ ion of m /z 402. cyanocinnamic acid (PCC). These two matrix molecules The detection limit (see Experimental Section) for this have more than one aromatic ring, and both possess adduct was found to be 275 fmol (see Table 1). The polar and nonpolar ends. Moreover, BCC and PCC detection limit, however, for the dibenzo[ a,[]pyrene- have strong UV absorption bands at 334 and 332 nm, 10-N ‘-guanine (DBP-N’-Gus) adduct was found to be respectively, which are nearly resonant with the wave- higher-2.5 pmol (see Table 1). Thus, the use of the length of the commonly used nitrogen laser. The dif- 4HCC matrix leads to variable detection limits for ference between the two matrices BCC and KC is the adducts depending on the hydrocarbon moiety. An- terminal or monosubstituted benzene ring. The inclu- other drawback associated with use of the 4HCC ma- sion of carbon and oxygen atoms to form BCC intro- trix for screening HPLC fractions from in vitro experi- duces additional degrees of freedom (i.e., rotations ments is that the PAH metabolites 3 and 4, which are about single bonds) into the molecule, and that gives two major components, are not detected with MALDI. more flexibility to BCC compared to PK. These matr- Moreover, the 4HCC matrix is not suitable for determi- ces were investigated more thoroughly than biphenyl nation of PAHs themselves because relatively large carboxylic acid and anthracene-9-carboxylic acid. amounts of analyte are required (see Table 1). A rea- The limits of detection for both ben&a]pyrene& sonable hypothesis is that these disadvantages can be N7-guanine (BP-6-N7-Gus), 1,and dibenzo[ a,1lpyrene- overcome by modifying the structure of IHCC. IO-N’-guanine (DBP-N’-Gua), 2, are in the femtomolar Biphenyl carboxylic acid, a molecule with two ben- range when 4-benzyloxy-a-cyanocinnamic acid (BCC) zene rings and lower polarity than 4-hydroxy-cl- is the matrix (see Table 1). Figure 2a and b demon- cyanocinnamic acid (4HCC), is expected to be a better strates that MALDI-TOF mass spectra of 2500 and 500 matrix for the metabolites. Metabolite 3 was desorbed fmol of DBP-N’-Gua, respectively, are obtained read- 1024 GEORGE ET AL. J Am Sot Mass Spectmm 1994,5,1021-1025

Table 1. Detection limits (ii femtomoles) for PAWDNA adduds, PAHs, and their metabolites Matrix 4HCC BCC PCC

BP-6-/W-Gua 11) 275 55 55 DBP-N7-Gus (2) 2500 500 200 DBP-i3,9-dial (3) n.d.’ n.d.’ 1000 DBP-11 ,l P-dial (4) n.d.a n.d.a 500 B[elP (6) 5000 500 500 DBP (6) 2000 250b 200

‘Not detectable ‘This sample was prepared as the others except approximately 50 ion-exchange resin beads (H+ form) were added to the wlution of matrix and analyte on the probe tip.

. a 1 a I

(M+H)+

b 402 ** 4w 500 * Figure 2. MALDI-TOF mass spectra of DBP-N’-Gua obtained & by using the BCC matrix. (a) 2500 fmol and (b) 500 fmol. The 20 &hi0 300 400 [M + I-II+ ion is of m/z 452. The asterisks (*I indicate matrix Figure 3. MALDI-TOF mass spectra of BP&N7-Gua obtained ions. by using the PCC matrix. (a) 11000 fmol and (b) 55 fmol. The (M + HI + ion is of m/z 402. The asterisks (“1 indicate matrix ions. ily by using BCC as the matrix. Metabolites 3 and 4, however, cannot be detected by using BCC matrix. The (see Table 1). The aromatic rings in KC are less PAHs, benzo[ elpyrene (B[ c]P) and dibenzo[ a,l]pyrene flexible than those of 4-benzyloxy-u-cyanocinnamic (DBP), were determined by using BCC. The detection acid (BCC), but more flexible than those of anthracene- limits are 500 and 250 fmol for B[e]P and DBP, respec- 9-carboxylic acid. Furthermore, KC is a good matrix tively (see Table l), and these limits are better than for detection of unreacted PAHs. The detection limits those obtained when 4HCC was used. for PAHs 5 and 6 are 500 and 200 fmol, respectively When 4-phenyl-cr-cyanocirmamic acid @CC>, which (see Table 1). resembles biphenyl carboxylic acid on one end and In summary, we have identified two new matrices, 4-hydroxy-or-cyanocinnamic acid (4HCC) on the other, 4-phenyl-cY-cyanocinnamic acid (PCC) and 4-betuyl- was employed as the matrix for the adducts 1 and 2, oxy-a-cyanocinnamic acid (BCC), for the determina- detection limits of 55 and 200 fmol, respectively, were tion of PAH-modified DNA bases at the femtomolar obtained (see Table 1). Figure 3a and b show the Level by MALDI-TOF. Additional improvements may MALDI-TOF mass spectra of 11,000 and 55 fmol of be achieved by incorporation of the recently developed benzo[a]pyrene-6-N7-guanine (BP&N7-Gual, respec- sample handling procedures of Vorm and Mann [28] tively, obtained by using FCC as the matrix. Metabo- and Xiang and Beavis [29]. The matrices developed Iites 3 and 4 of molecular weight 336 can be detected here are also suitable for the detection of the PAHs as radical cations by using this matrix. The detection themselves. We also wish to advocate the concept of limits for 3 and 4 are 1000 and 500 tinol, respectively designing the MALDI matrix to have structural simi- J Am Sue Mass Spectrom 1994,5, 1021~1025 MATRIX DESIGN FOR MALDI 1025

larities to the particular analyte rather than selecting 14. Danis, P. 0.; Kar, D. E. Org. Mass Spectrom 1993, 28, 923-925. vn an empirical basis. 15. Bahr, U.; Deppe, A.; Karas, M.; Hillenkamp, F.; Giessman, U. Anal. Chem. 1992, 64, 2866-2869. 16. Stahl, B.; Steup, M.; Karas, M.; Hillenkamp, F. Anal. Chem. Acknowledgment 1991, 63, 1463%1466. This work was supported by the National Institutes of Health 17. Tang, K.; Allman, S. L.; Jones, R. B.; Chen, C. H.; Araghi, S. (Grant No. lPOlCA49220). Rapid Commun. Muss Spectrom. 1993, 7, 63-66. 18. Wang, B. H.; Biemann, K. Anal. Chem. 1994, 66, 1918~1924. 19. Wu, K. J.; Shaler, T. A.; Becker, C. H. Anal. Chin. 1994, 66, References 1637-1645. 20. Lay, J. 0. Jr.; Chiarelli, M. P.; Bryant, M. S.; Nelson, R. W. 1. luhasz, P.; Costello, C. E.; Biemann, K. J. Am. SK. Mass Enumm. Health Pwspewcr.1993, 99, 191-194. Spcc~ctrom.1993, 4, 399-409. 21. Hettich, R.; Buchanan, M. 1. Am. Sec. Mass Spectrom. 1991, 2, 2. Philips, D. H.; Sims, P. In Chemical Carcirqens and DNA; 402-412. Grover, P. L., Ed.; CRC Press: Boca Raton, 1979; p 29. 22. Stemmler, E. A.; Buchanan. M. V.; Hurst, G. 8.; Hettich, R. L. 3. Sims, I’.; Grover, P. L.; Swaisland, A.; Pal, K.; Hewer, A. Anal. Chew. 1994, 66, 1274-1285. Nahrre 1974, 252, 326-327. 23. RamaKrishna, N. V. 5.; Padmavathi, N. S.; Cavalieri, E. L.; 4. Rogan, E. G.; Cavalieri, E. L. In Free Radicals in Bioloa; Pryor, Rogan, E. G.; Cerny, R. L.; Gross, M. L. Chum. Res. Toxicol. W. A., Ed.; Academic Press: New York, 1984; p 323. 1993, 6, 554-562. 5. Stansbuty, K. H.; Flesher, J. W.; Gupta, R. C. Chem. Res. Toxicol. 1994, 7, 254-259. 24. Child, R. C.; Osterberg, A. C.; Sloboda, A. E.: Tomcufcik, A. 6. We&mans, J.; Cerny, R. L.; Gross, M. L. The Analyst, 1994, S. J. Pharmactrl. Sci. 1977, 66, 466-476. 119, 497-503. 25. Rogan, E. G.; Cavalieri, E. L.; Tibbels, S. R.; Cremonesi, I’.; 7. Crain, P. F. Mass Spectrom. Rev. 1990, 9, 505-554. Warner, C. D.; Nagel, D. L.; Tamer, K. B.; Gross, M. L. J. Am. 8. Chiarelli, M. I’.; Lay, J. O., Jr. Mass Spectrom. Rev. 1992, 12, Chem. SOC.1988, 110, 4023-4029. 447-493. 26. RamaKrishna, N. V. S.; Padmavathi, N. S.; Cavalieri, E. L.; 9. McCloskey, J. A.; Grain, P. F. Int. J. Mass Spectvom. ion Rogan, E. G.; Cemy, R. L.; Gross, M. L. Chem. Res. Toxicol. Processes 1992, 118/119, 593-615. 1993, 6, 554-560. 10. Wellemans, J.; George, M.; Cemy, R. L.; Cross, M. L. Pooly- 27. Cavalieri, E. L.; Rogan, E. G.; Devanesan, P. D.; Cremonesi, cyclic Aromatic Compounds 1994, 5, in press. I’.; Cerny, R. L.; Gross, M. L.; Bodell, W. J. Biochemistry 1990, 11. Chait, 8. T.; Wang, R.; Beavis, R. C.; Kent, S. B. Science 1993, 29, 4820&4827. 262,8Y-92. 28. Vorm, 0.; Roepstorff, P.; Mann, M. Proceedings of the 42nd 12. Karas, M.; Ingcndoh, A.; Bahr, U.; Hillenkamp, F. Biomed. ASMS Conference on Mass Specctrometry and Allied Topics; Emiron. Mnss Specctrom. 1969, 18, 84-843. Chicago, IL, May 29-June 3, 1994, p 9. 13. Danis, P. 0.; Ku, D. E.; Mayer, F.; Holle, A.; Watson, C. H. 29, Xiang, F.; Beavis, R. C. Rdptd Commun. MOSSSpectvom. 1994, Org. Mass Spectrom. 1992, 27, 843-846. 8,199-204.